The data types

What data AI labs license from companies, in the buyers' words

Updated

Every buyer on the table names what it takes, and the lists fall into a few families of record.

Check my company's records Send the size of the company, the systems the records sit in and how many years they cover. We reply with the buyers whose stated asks fit and what has to come out first.

Records of how a company runs

  • Nyne: CRM, HR, ERP and support records from core systems.
  • Maryah: emails, sales records and support tickets a company chooses to share.
  • micro1: SOPs, knowledge bases, internal documentation, CRM data, project histories and QA processes.
  • Grepped: messages, docs, tickets, code and records.
  • Mercor: operational data for frontier labs.

Source: AI data buyer requirements, by buyer

Code, media and books

SimpleClosure asks for codebases, repositories and documentation; Wirestock for photography, video, illustration and design; Created by Humans for books; Protege lists healthcare, video, audio and speech, and spatial data among its domains.

Personal data

7 of 12 say they remove or de-identify personal data, and 1 asks the seller to leave it out altogether.

Source: AI data buyer requirements, by buyer

Check which buyers fit your company's records

Has a buyer already made an offer for the data?

Tick the box above to send your enquiry.

Sources

This site's own dataset, every figure quoted from its source and dated:

Your records match a buyer's ask?

Send the size of the company, the systems the records sit in and how many years they cover. We reply with the buyers whose stated asks fit and what has to come out first.

Check my company's records